COS 165-3 - Mixed signals: Innately attractive and learned volatile cues mediate bumblebee preferences for Mimulus guttatus flowers

Thursday, August 10, 2017: 2:10 PM
C125-126, Oregon Convention Center
Ariela I. Haber1, James W. Sims2, Mark C. Mescher2, Consuelo M. De Moraes2 and David E. Carr3, (1)Department of Environmental Sciences, University of Virginia, Charlottesville, VA, (2)Department of Environmental Systems Science, ETH Zurich, (3)Blandy Experimental Farm, University of Virginia, Boyce, VA

Floral scent is a key trait that plants use to advertise rewards. Associative learning in pollinators is expected to select for production of honest signals of floral rewards in plants. However, evolution of and contraints on honest signaling are not well-studied. Previous work in the mixed-mating, bumblebee-pollinated herb Mimulus guttatus demonstrated that bumblebees prefer outbred over inbred plants, but this preference is independent of reward quality. Inbreeding in M. guttatus reduces emission of β-trans-bergamotene, but β-trans-bergamotene is not correlated with rewards. However, six compounds that are unaffected by inbreeding are positively correlated with rewards. We hypothesized that compounds that are correlated with rewards function as honest signals in M. guttatus, but that β-trans-bergamotene interferes with these signals. We conducted a series of pairwise choice tests using a Y-tube and using artificial flowers in a flight cage to determine whether 1) bumblebees are innately attracted to β-trans-bergamotene, 2)bumblebees develop a preference for compounds that are correlated with rewards, and 3) β-trans-bergamotene interferes with the ability for bumblebees to discriminate based on honest signals. To address objectives 2 and 3, we compared preferences for limonene, which was positively correlated with rewards, and linalool, which was slightly negatively correlated with rewards.


Bees demonstrated an innate preference for β-trans-bergamotene in pairwise choice tests using live plants and using extracted floral volatile samples. Bumblebees did not innately prefer limonene or linalool. After foraging on live M. guttatus flowers with varying reward quality, bumblebees demonstrated a significant preference for limonene, but not linalool, indicating that bumblebees can form preferences for M. guttatus floral odors through associative learning. However, bumblebees only exhibited this learned preference when odor choices were presented over a background of M. guttatus floral scent that did not contain β-trans-bergamotene. When β-trans-bergamotene was present in the background, bumblebees did not exhibit any preference. Thus, β-trans-bergamotene interferes with bumblebees’ ability to make choices based on honest signals. We suspect that its innately attractive properties mediate bumblebee behavior in a way that is analogous to pheromone-mimicking compounds in deceptive orchids. Such deceptive signaling mechanisms have not previously been demonstrated in flowers that also offer rewards to pollinators. Perhaps reducing the covariance between signal and reward allows plants to minimize costs of reward production without causing avoidance learning in pollinators. These results demonstrate that volatile signals from plants to pollinators can be highly complex, and suggest that the evolution of honest signals is not straightforward.